COST-EFFECTIVENESS OF ARTIFICIAL INTELLIGENCE (AI)-ASSISTED BRAIN MAGNETIC RESONANCE IMAGING (MRI) INTERPRETATION IN ADVANCED LUNG CANCERS
Author(s)
Hsiao Ling Chen, MS1, Chen-Han Chueh, PhD2, Ming-Yu Hong, MS3, Jia-Sheng Hong, MS4, Chien-Yu Tseng, PharmD5, Yu-Te Wu, PhD6, Wan-Yuo Guo, PhD7, Shuu-Jiun Wang, PhD7, Yi-Wen Tsai, PhD8.
1Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taipei, Taiwan, 2University of California San Diego, San Diego, CA, USA, 3Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taiwan, New Taipei City, Taiwan, 4Institute of Biophotonics National Yang Ming Chiao Tung University Taipei, Taiwan, Taipei, Taiwan, 5Department of Pharmaceutical Outcomes and Policy, College of Pharmacy, University of Florida, Gainesville, FL, USA, 6Institute of Biophotonics, National Yang Ming Chiao Tung University, Taipei, Taiwan, 7College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan, Taipei, Taiwan, 8National Yang Ming Chiao Tung University, Taipei, Taiwan.
1Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taipei, Taiwan, 2University of California San Diego, San Diego, CA, USA, 3Institute of Health and Welfare Policy, National Yang Ming Chiao Tung University, Taiwan, New Taipei City, Taiwan, 4Institute of Biophotonics National Yang Ming Chiao Tung University Taipei, Taiwan, Taipei, Taiwan, 5Department of Pharmaceutical Outcomes and Policy, College of Pharmacy, University of Florida, Gainesville, FL, USA, 6Institute of Biophotonics, National Yang Ming Chiao Tung University, Taipei, Taiwan, 7College of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan, Taipei, Taiwan, 8National Yang Ming Chiao Tung University, Taipei, Taiwan.
OBJECTIVES: Brain MRI is routinely performed in patients with advanced non-small cell lung cancer (NSCLC) for identifying brain metastases (BM). Accurate MRI image interpretation directly influences timely cancer treatment strategies. This study evaluated the cost-effectiveness of DeepBT®, an AI -based decision-support SaMD for brain MRI interpretation in stage III-IV NSCLC from the Taiwan National Health Insurance perspective.
METHODS: A four-state Markov model (no brain metastasis, brain metastasis, end-of-life supportive care, and death) simulated clinical pathways following brain MRI interpretation by radiologists with and without DeepBT® over a 10-year horizon, using 1-year cycles and a 3% discount rate. Improved diagnostic accuracy with DeepBT® was assumed to influence health state transitions, while false-positive and false-negative interpretations were associated with unnecessary costs, delayed or inappropriate treatment, disease progression, and quality-of-life loss. Outcomes included life-years, quality-adjusted life-years (QALYs), total costs, and incremental cost-effectiveness ratios (ICERs), evaluated against a willingness-to-pay threshold (WTP) of twice Taiwan’s 2024 GDP per capita (NTD 2.336 million). Parameter uncertainty was assessed through deterministic and probabilistic sensitivity analyses (DSA and PSA).
RESULTS: The ICER for DeepBT®-assisted MRI interpretation was national Taiwan dollar (NTD) 2,243,160 per QALY, below the WTP threshold. PSA demonstrated a 99.2% probability of cost-effectiveness despite higher total costs, and DSA indicated that results were most sensitive to health utility values and costs associated with brain metastases. Scenario analyses varying transition probabilities, time horizon, and discount rate consistently supported the base-case cost-effectiveness conclusion.
CONCLUSIONS: DeepBT®-supported brain MRI interpretation represents a cost-effective diagnostic strategy for patients with advanced NSCLC, with relevance for reimbursement decisions and routine clinical practice.
METHODS: A four-state Markov model (no brain metastasis, brain metastasis, end-of-life supportive care, and death) simulated clinical pathways following brain MRI interpretation by radiologists with and without DeepBT® over a 10-year horizon, using 1-year cycles and a 3% discount rate. Improved diagnostic accuracy with DeepBT® was assumed to influence health state transitions, while false-positive and false-negative interpretations were associated with unnecessary costs, delayed or inappropriate treatment, disease progression, and quality-of-life loss. Outcomes included life-years, quality-adjusted life-years (QALYs), total costs, and incremental cost-effectiveness ratios (ICERs), evaluated against a willingness-to-pay threshold (WTP) of twice Taiwan’s 2024 GDP per capita (NTD 2.336 million). Parameter uncertainty was assessed through deterministic and probabilistic sensitivity analyses (DSA and PSA).
RESULTS: The ICER for DeepBT®-assisted MRI interpretation was national Taiwan dollar (NTD) 2,243,160 per QALY, below the WTP threshold. PSA demonstrated a 99.2% probability of cost-effectiveness despite higher total costs, and DSA indicated that results were most sensitive to health utility values and costs associated with brain metastases. Scenario analyses varying transition probabilities, time horizon, and discount rate consistently supported the base-case cost-effectiveness conclusion.
CONCLUSIONS: DeepBT®-supported brain MRI interpretation represents a cost-effective diagnostic strategy for patients with advanced NSCLC, with relevance for reimbursement decisions and routine clinical practice.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
EE45
Topic
Economic Evaluation
Disease
SDC: Oncology